Preparing the Data
The collected student pre/post-survey data were linked with student
cumulative GPA, ACT scores, and demographics data (sex, ethnicity,
year in school, age, and major). Data were analyzed only for students
who gave consent, answered the control question correctly, and completed both surveys in their entirety. Of the 130 students in the class,
112 met these criteria. IMI items 11, 12, 21, 22, and 24 needed to be
reversed before subscale totals could be calculated; this was done as
previously described (Ryan, 1994). A 0– 6 Likert scale was used for all
items so that a completely unmotivated student would receive a zero.
Aggregated variables were created representing four IMI and
four MSLQ subscales mentioned above. Student data for each subscale were aggregated, averaged, and converted to a percentage for
ease of interpretation. Many students were acutely aware of their
course performance at the time the post-survey was conducted,
so the perceived confidence (IMI) and self-efficacy for learning
and performance (MSLQ) subscales were left out to avoid confounding this analysis.
Cronbach’s alpha statistics were calculated for each of the IMI and
MSLQ dimensions (Table 2). These Cronbach alpha values indicate
that the internal consistency within these scales ranged from
acceptable (0.8 > ɑ ≥ 0.7) to excellent (ɑ ≥ 0.9).
IMI & MSLQ Pre/Post-Survey Results
Student pre/post-survey scores for the eight subscales used in this
study are shown in Figure 1. Scores on the pre-survey ranged from
43% on the anxiety subscale (MSLQ) to 82% on the value subscales
(MSLQ and IMI). High pre-survey scores were predicted for these stu-
dents, as this class is the first biology class they take for their major
and the students are typically very enthusiastic when they enter the
course (B. Gibbens, personal observation). Despite this, student scores
on the intrinsic motivation, self-efficacy, and value subscales declined
during the semester. The pre/post-survey scores and the amount of
decline were both very similar between comparable MSLQ and IMI
subscales (Figure 1). For example, student self-efficacy scores declined
by 13% on the IMI and by 12% on the MSLQ.
An ordinary least squares (OLS) regression analysis was conducted to
determine which motivation scores, if any, predicted overall student
performance in the class. Each student’s final course percentage was
used as a metric for their overall course performance. The regression
results show that value/usefulness (IMI), pressure/tension (IMI), and
test anxiety (MSLQ) pre-survey scores were the best predictors of
course performance (Figure 2). Scores on the value/usefulness (IMI)
subscale positively predicted course performance, whereas scores on
the pressure/tension (IMI) and text anxiety (MSLQ) subscales negatively predicted performance.
Separate regression models were constructed to explore the relationship between overall course performance and changes in levels
Table 1. Example items for each subscale, taken verbatim from the original Intrinsic Motivation Inventory
(IMI) and Motivated Strategies for Learning Questionnaire (MSLQ) surveys, with total number of items (n)
asked from that subscale on the pre/post-surveys.
Example n Example n
Intrinsic I enjoyed doing this activity very
7 The most satisfying thing for me in this
course is trying to understand the
content as thoroughly as possible.
Self-efficacy I think I am pretty good at this
6 I’m confident I can do an excellent job
on the assignments and tests in this
Value I think this is an important activity. 4 It is important for me to learn the course
material in this class.
Anxiety I felt very tense while doing this
5 I feel my heart beating fast when I take
Table 2. Cronbach’s alpha statistics for each of the
Intrinsic Motivation Inventory (IMI) and the
Motivated Strategies for Learning Questionnaire
Instrument Subscale Cronbach’s ɑ
MSLQ Intrinsic goal
Task value 0.885
Test anxiety 0.717